Surgical site infection (SSI) represents a major burden for patients, doctors and health systems around the world, but is potentially preventable. SSI is the most common postoperative complication across all income and development settings, and the most common healthcare-associated infection in low-and middle-income countries (LMICs). It has been associated with one-third of postoperative deaths and accounts for 8% of all deaths caused by a nosocomial infection. Rates vary significantly between different types of surgery, but it is particularly prevalent in abdominal operations; as many as one in four patients get an SSI when the operation involves the bowel.
Improving surgical outcomes is a global health priority, highlighted by the Lancet Commission on Global Surgery. Recent WHO guidelines made 29 recommendations for intraoperative and postoperative measures to prevent SSI. Despite inclusion of strongly graded recommendations, there was little high-quality evidence in support of most interventions. In addition, none of the evidence used was derived from resource limited settings, leading to uncertainty about implementation of measures in LMICs, in particular. The specific interventions to be tested in this study were selected by a Delphi process from a long list of potential interventions based on the the WHO guidelines.
A key factor influencing SSI rates is intraoperative contamination (opening the gut, genital or urinary tracts). In clean (non-contaminated) surgery, the infection rate is low and benefits from research targeting SSI rates in those patients are limited. CHEETAH will therefore focus on clean-contaminated, contaminated and dirty surgery. A cluster randomised design is considered the most appropriate design to minimise the risk of contamination in the control arm of the trial. Once theatre staff have been trained to implement the intervention (of separating sterile surgical instruments and gloves at the time of wound closure) it is likely to become routine practice, making individual randomisation unfeasible.